llvm/torch-mlir
A compiler that translates PyTorch models into MLIR dialects for hardware deployment.
Not currently ranked — collecting fresh signals.
star history
Torch-MLIR bridges the PyTorch ML ecosystem to the MLIR compiler infrastructure, enabling hardware vendors to lower PyTorch models into MLIR and subsequently to their target backends. It provides multiple lowering paths from PyTorch through ONNX and FX entry points, similar to how LLVM supports multiple hardware targets through a common intermediate representation.
Frequently asked
- What is llvm/torch-mlir?
- A compiler that translates PyTorch models into MLIR dialects for hardware deployment.
- Is torch-mlir open source?
- Yes — llvm/torch-mlir is an open-source project tracked on heatdrop.
- What language is torch-mlir written in?
- llvm/torch-mlir is primarily written in C++.
- How popular is torch-mlir?
- llvm/torch-mlir has 1.9k stars on GitHub.
- Where can I find torch-mlir?
- llvm/torch-mlir is on GitHub at https://github.com/llvm/torch-mlir.